Questions on Ryu algorithm

[zejal]

I am trying to take a grasp on how Ryu algorithm reading your publication in 2018 (https://doi.org/10.1145/3192366.3192369) and reading also code you published in GitHub.

There a couple of things I do not understand from the article related to how q is set. In lemma 3.2, it is mentionned that q is set to :
For e2 < 0, choose q = ⌊−e2 log10 5⌋.
But later on in lemma 3.4 and steps 0 and 3’, article reads : q as max (0, ⌊−e2 log5 2⌋ − 1)

Ryu implementation also uses log10 5

// This expression is slightly faster than max(0, log10Pow5(-e2) - 1).
const uint32_t q = log10Pow5(-e2) - (-e2 > 1);

so it is not clear to me why there is this difference between article and implementation.

Rather minor, it seems to me that captions for 2 figures in left column in page 278 are inverted.